dplyr mutate, replace, and column lengths

Question

library(tidyverse)
df <- tibble(col1 = c("A", "B", "C", "D"),
             col2 = c(NA, NaN, Inf, 5),
             col3 = c(NA, NaN, Inf, 5))
#> # A tibble: 4 x 3
#>   col1   col2  col3
#>   <chr> <dbl> <dbl>
#> 1 A        NA    NA
#> 2 B       NaN   NaN
#> 3 C       Inf   Inf
#> 4 D         5     5

I'd like to replace col2 column values NA and NaN with 0. I'm aware I could do this with the replace_na() function, but instead am trying it in the following manner... and getting this column length error.

df %>% mutate(col2 = replace(col2, is.na(.), 0))
#> Error: Column `col2` must be length 4 (the number of rows) or one, not 12

col2 is indeed four rows in length, not twelve. Is the error misleading or am I misinterpreting something? And what to do to correct it?

Answer 1

The problem is that the . in your statement is referring to df (which has 12 elements), not to col2. You need to write is.na(col2) explicitly...

df %>% mutate(col2 = replace(col2, is.na(col2), 0))

Alternatively you can use replace_na, which is designed just for this purpose...

df %>% mutate(col2 = replace_na(col2, 0))